Download Fetal Hemoglobin Levels in Sickle Cell Disease and

From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
Fetal Hemoglobin Levels in Sickle Cell Disease and Normal Individuals Are
Partially Controlled by an X-Linked Gene Located at Xp22.2
By G.J. Dover, K.D. Smith, Y.C. Chang,
S.Purvis, A. Mays, D.A. Meyers, C. Sheik, and G. Serjeant
Fetal hemoglobin (Hb F) production in sickle cell (SS) disease
and in normal individuals varies over a 20-fold range and is
under genetic control. Previous studies suggested that variant Hb F levels might be controlled by genetic loci separate
from the pglobin complex on chromosome 11. Using microscopic radial immunodiffusion and flow cytometric immunofluorescent assays to determinethe percentage of F reticulocytes and F cells in SS and nonanemic individuals, we
observed that F-cell levels were significantly higher in nonanemic females than males (mean f SD, 3.8% f 3.2% Y
2.7% f 2.3%). F-cell production as determined by F reticulocyte levels in SS females was also higher than in SS males
(17% f 10% v 13% f 8%). We tested the hypothesis that
F-cell production in both normal and anemic SS individuals
was controlled by an X-linked locus with two alleles, high (H)
and low (L). Using an algorithm to determine the 99.8%
confidence interval of a normal distribution in nonanemic
individuals, we estimated that males and females with at
least one H allele had greater than 3.3% F cells. Comparisons
of male-male or female-femaleSS sib pairs with discordant F
reticulocyte levels distinguishedtwo phenotypes in SS males
(L, ~ 1 2 % ;H, >12%) and three phenotypes in SS females
(LL, ~ 1 2 % ;HL, 12% to 24%. HH >24%). Linkage analysis
using polymorphic restrictionsites along the X chromosome
in eight SS and one AA family localized the F-cell production
(FCP) locus to Xp22.2, with a maximum lod score (logarithm
of odds of linkage v independent assortment) of 4.6 at a
recombination fraction of 0.04.
0 1992 by The American Society of Hematology.
P
and, in the case of SS disease, the preferential survival in
the peripheral circulation of F cells compared with RBCs
containing no detectable Hb F.9 In SS individuals, the
variation in F-cell production as measured by the percent F
reticulocytes is the major variable contributing to differences in Hb F levels.
The broad distribution of Hb F levels in normal adults
and in anemic SS individuals suggests that more than one
genetic factor may control Hb F production. Several point
mutations in the promoter regions of the two gamma genes
have been associated with increased Hb F production in
otherwise normal adults, thereby indicating that factors
linked to the P-globin gene complex on chromosome 11
influence Hb F production. Differences in mean Hb F levels
in SS individuals homozygous for each of the three most
common P-globin gene restriction fragment length polymorphism (RFLP) haplotypes also suggest that genetic factors
linked to chromosome 11influence Hb F levels.loJ1
Several laboratories have suggested that a locus controlling F-cell production (FCP locus) in normal individuals
and in SS disease may be separate from the P-globin gene
complex located on chromosome 11. Five families have
been described in which heterocellular HPFH was not
linked to the P-globin gene complex.12-16Comparison of
F-cell production between SS siblings indicated that one
third of the sib pairs had different levels of F-cell production, thus indicating that at least one FCP locus in SS
disease is not linked to the P-globin gene c0mp1ex.l~
Recently, Miyoshi et a1 reported evidence suggesting that
F-cell production in heterocellular HPFH within the Japanese population is controlled by an X-linked locus.1s
Thus, Hb F production is a controlled by multiple genetic
factors both linked and unlinked to the P-globin gene
complex on chromosome 11. In this report, we have
attempted to localize one of the unlinked factors: a gene
that controls F-cell production in SS and normal individuals. Our initial analysis tests the simplest genetic model
consistent with our analysis of F-cell levels. This model is
based on the assumption that a single locus responsible for
a major portion of thevariation in a quantitative trait can be
genetically mapped by linkage to polymorphic DNA loci.l9
OSTNATALLY, fetal hemoglobin (Hb F) production
persists and is confined to a subset of erythrocytes,
termed F cells.’ ChemofF and Marti and Butler3 demonstrated that Hb F levels in normal adults vary over a 20- to
35-fold range. Hb F levels in individuals with sickle cell
anemia (SS), albeit higher than in normal adults, also vary
over a 20-fold range? Marti and Butler? using lysate levels
of Hb F, and Zag0 et a1,5 using immunologic techniques for
counting F cells, showed that high Hb F (1.1% to 3.4%) and
high F-cell levels (> 8%) in normal adults were inherited.
The terms applied to these examples of hereditary persistence of Hb F (HPFH) were “Swiss type HPFH’ or
heterocellular HPFH.3,5Dover et aI6 and later Mason et a17
and Milner et aI8showed that either F-cell or Hb F levels in
nonanemic sickle trait (AS) parents predicted Hb F production in their SS children. The genetic mechanisms underlying this variation in Hb F production in normal and anemic
individuals are unclear.
Because Hb F is unevenly distributed within the red
blood cells (RBCs) of normal and SS individuals, differences in peripheral blood Hb F levels may be attributed to
at least three distinct variables: differences in the production of F cells, variation in the amount of Hb F per F cell,
From the Department of Pediatrics, Medicine, The Johns Hopkins
University School of Medicine, and the Kennedy Institute, Baltimore,
MD; and the MRC Sickle Cell Center, University of the West Indies,
Kingston, Jamaica.
Submitted September 27, 1990; accepted April 14,1992.
Supported by Grant No. HL28028 from the National Heart Lung
and Blood Institute, National Institutes of Health (NIH), Bethesda,
MD; and the MRC Laboratories, Kingston, Jamaica. Y.C. Chang is
supported by NIH Training Grant No. 5T32GM10714-09.
Address reprint requests to George J. Dover, MD, Ross Research
Building, Room 1125, Johns Hopkins University Medical School, 720
Rutland Ave, Baltimore, MD 21205.
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked
“advertisement” in accordance with 18 U.S.C. section 1734 solely to
indicate this fact.
0 1992 by The American Society of Hematology.
0006-4971/92/8003-0027$3.00/0
816
Blood, VOI 80, NO3 (August 1). 1992: pp 816-824
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
ADULT Hb F LEVELS ARE INFLUENCED BY AN Xp GENE
In this report, we provide evidence for a FCP locus that is
linked to the short arm of the X-chromosome within the
Xp22.2 region.
MATERIALS AND METHODS
Subjects. Venous blood was collected in either heparinized or
EDTA-containing tubes after verbal or written informed consent
as approved by the Johns Hopkins University Medical School Joint
Committee on Clinical Investigation. Normal adult samples were
obtained from blood donors at Johns Hopkins and from normal
controls used in a previously published survey.20All donors had
normal hemoglobin levels and normal RBC indices. Pregnant
females or recently pregnant females ( < 6 months since delivery)
were excluded. SS subjects and their families were identified from
the clinic population at Johns Hopkins or from the Medical
Research Council Sickle Cell clinic at the University of West
Indies, Kingston, Jamaica. Two groups of SS patients from Jamaica
were studied. A sib pair group was derived from 81 families with
more than one child with SS disease. Part of this group was the
subject of a previous publicati~n,'~
which outlined criteria for the
diagnosis of SS disease, reproducibility of F-cell determinations,
and validation of paternity. A second group, termed the cohort
group, included 103 subjects chosen randomly from 275 SS subjects
identified in a prospective neonatal screening program in Jamaica.21 The age range for the sib pair group was 5 to 50 years
(mean ? SD, 19.6 ? 12.7 years) and 5 to 15 years for the cohort
group (9.8 +. 2.3 years). All subjects in Jamaica have been monitored for more than 5 years and their steady-state hematologic tests
have been repeatedly assayed. A white family with heterocellular
hereditary persistence of Hb F and X-linked adrenoleukodystropy
was also studied.22 Iron-deficiency anemia was excluded in the
nonanemic subjects on the basis of a normal hemoglobin for age
and normal RBC indices. In SS subjects, elevated serum ferritin
excluded iron deficiency.
F-cell assays. Assays for percent F cells were performed on
peripheral blood to determine F-cell production in nonanemic
normal (Hb AA) and AS subjects. Monoclonal mouse anti-human
Hb F antibody was used to enumerate lo4 RBCs per assay using a
flow cytometry technique.z3 An assay for the percent of reticulocytes containing Hb F was used to measure F-cell production in SS
subjects.24We have previously documented that due to preferential
survival of F cells in SS subjects, the percent F reticulocytes is the
most accurate measurement of F-cell production in SS subject^.^
DNA analysis. The haplotypes resulting from the pattern of
RFLPs within the P-globin gene complex were determined for all
SS subjects to help exclude nonpaternity in SS sib pairs.2s RFLP
analysis in all subjects used in linkage studies was performed on
DNA derived from whole blood or Epstein-Barr virus (EBV)transformed lymphocytes using X-chromosome probes obtained
from the ATCC (Rockville, MD), Collaborative Research (Bedford, MA), or individual investigators. Probes used for linkage
analysis included DXS 52 (Xq28), HPRT (Xq26), DXS 11 (Xq2524), DXS 17 (Xq22), PGK (Xq13), DXS 14 (Xp11.21), DXS 7
(Xp11.4-11.3), DXS 84 (Xp21.1), DXS 67 (Xp21.3), DXS 41
(Xp22.1), DXS 274 (Xp22.2-22.1), DXS 16 (Xp22.2), DXS 43
(Xp22.2), DXS 85 (Xp22.2-22.3), DXS 143 (Xp22.3), and DXS 452
(Xp22.3).
Linkage analysis. Linkage analysis was performed using the
computer program LIPED (courtesy of J. Ott, Columbia University
School of Medicine, New York, 1987). This program uses the
method of maximum likelihood to calculate lod scores (logarithm
of odds of linkage v independent assortment) at selected recombination fractions for each pedigree, and permits summing lod scores
from individual pedigrees. Because no single probe was informa-
817
tive in all families, multiple probes from the same region were
tested. For determination of lod score, only one informative probe
per region was used with each pedigree. Linkage analysis was
performed by treating the variation in F-cell production as either a
discrete or a continuous trait. The criteria used for defining FCP
phenotypes are described in the Results. For the continuous trait
analysis, the mean r SD of FCP phenotypes were defined as
follows: high (HH) 31.8% ? 6.9% F reticulocytes, intermediate
(HL) 16.5% ? 3.5% F reticulocytes, and low (LL) 6.2% ? 3.0% F
reticulocytes. Complete penetrance was assumed for each allele.
Allele frequencies were derived from analysis of the cohort group.
Where indicated, we analyzed our maximum lod score as described
by Ott,'9,26927for situations where the mode of inheritance is
unknown.
RESULTS
F-cell production in normal and A S adults. Using our
flow cytometry technique for measuring F-cell levels, we
examined 292 adults (Fig 1).We found that females have a
significantly higher percentage of F cells than males
(3.8% +- 3.0%, female; 2.7% 2.3%, males; P = .003).
Mason et al,' studying children less than 6 years old, and
Rutland et a1,28studying adults, have previously described
higher Hb F levels in nonanemic females than males. The
*
Fig 1. F-cell production in nonanemic adults. The percent F cells
was determined by flow cytometric assays of 10,000 cells per assay
using monoclonal mouse anti-human Hb F antibody.z3 The skewed
distribution is identical to that seen by Miyoshi et aL18 Results for 292
healthy adults (A, 121 males; B, 171 females) are shown. The broken
vertical line (3.3% F cells) is the 99.8% confidence interval for a best-fit
normal distribution (P= .14; mean f SD for the calculated normal
distribution, 1.7 f 0.7).
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
DOVER ET AL
818
skewed distribution and the range of F cells in our study
(0.2% to 20%) are similar to that observed in the Japanese
adults studied by Miyoshi et all8 ( < 1% to 16%) and Zag0
et a15 (< 1% to 18%). Black and white adults had similar
average percent F cells (3.3% 2 2.9% and 3.0% 2 2.7%,
respectively). The range for 25 AS individuals (0.2% to
15%) is identical to that seen in AS subjects studied by
Wood et al?9
F-cell production in SS subjects. Because of preferential
survival of F cells in SS subjects, the most accurate estimate
of F-cell production in individuals with SS is the F-reticulocyte count? In both the sib pair and the cohort groups of S S
patients, the percent F reticulocytes in females was higher
than males (sib group: females, 18% & 11% [n = 981 v
males, 14% ? 9% [n = 801, P = .02;cohort group: females,
13% 2 7% [n = 451 v males, 10% 2 7% [n = 581, P = .006).
In the case of S S sib pairs, both individuals in a given sib
pair must have identical loci at the @-globingene complex.
Thus, the difference in F-cell production must be due to
genetic factors elsewhere. This difference was evident at all
age levels, suggesting that iron deficiency, which is more
prevalent in menstruating females, is not a cause for the
difference in F-cell production. The percent F reticulocytes
was not correlated with either the hemoglobin level (g/dL)
or the percent total reticulocytes. A striking difference in
the distribution of percent F reticulocyte is seen when S S
males are compared with S S females (Fig 2). While SS
males have a skewed unimodal distribution, S S females
have a distribution that appears to be bimodal. The limited
number of females analyzed did not allow statistical confirmation of a bimodal distribution; however, the probability
of bimodality was 80% by the bimodality test of Giacomelli
et a1.M
The discrepancy between male and female F-reticulocyte
levels suggested the possibility that gender differences
might have falsely elevated the proportion of S S sib pairs
with discordant F-reticulocyte levels in the previous SS sib
study.17 We had previously defined F-reticulocyte levels
between S S sibs as being discordant if the probability that
F-reticulocyte levels for S S sibs were alike was
as
determined by Student's t-test. Therefore, we reexamined
our now expanded sib pair study group to determine if
differences in F-cell production among SS sibs were related
to gender. Among 97 sib pairs, 39 (40.2%) had significantly
different percent F reticulocyte levels. Among sibs of the
same sex, 15 of 45 (33%) had discordant percent F
reticulocytes. Thus, gender differences between sib pairs
cannot explain the differences in F-cell production between
all S S sibs. Since these S S sibs must have identical @-globin
loci, this reanalysis confirms our previous finding" that one
control of F-cell production in S S disease is separate from
the @-globinlocus.
X-linked FCP locus. It is clear that there are marked
differences in the number, range, and distribution of F-cell
production in males and females. While sex-limited expression could explain these differences, the simplest hypothesis is that the responsible genetic locus is on the X-chromosome. Therefore, we tested the hypothesis that a single FCP
12
10
8
6
0
10
20
30
40
0
10
20
30
40
2
0
% F RETICULOCYTES
Fig 2. Percent F reticulocytes in SS cohort patients. The percent F
reticulocytes was determined by radial immunodiffusion assay.=
Results from 103 SS subjects (A, 58 males; 6.45 females) selected
randomlyfrom a cohort of SS patients identifiedat birthn are shown.
locus with two alleles, high (H) and low (L), exists on the
X-chromosome.
Phenotypes and genotypes of nonanemic adults. Because
nonanemic adults have a skewed unimodal distribution of
percent F cells, it has been difficult to distinguish high or
low FCP phenotypes. However, Miyoshi et a1 were able to
distinguish high and low F-cell groups using a least-squares
method.18 An algorithm that sequentially eliminates the
highest values one by one was used to determine the best fit
for a normal distribution of the total data. A 99% confidence interval for the normal distribution was used to
distinguish the high and low phenotypes. In Miyoshi's
analysis, using microscopic examination of blood smears
treated with a rabbit anti-human Hb F antibody, high F-cell
production was set at greater than 4.4%.Applying the same
algorithm to our flow cytometry-determined percent F
cells, we identified a high group beginning at 3.3% (Fig 1).
This analysis allows assignment of high (H) and low (L)
phenotypes for males and females. Based on the hypothesis
of a X-linked gene, we can deduce that males have either an
H allele ( > 3.3% F cells) or an L allele ( < 3.3% F cells). In
females, we can infer the LL genotype ( < 3.3% F cells), but
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
ADULT Hb F LEVELS ARE INFLUENCED BY AN Xp GENE
cannot distinguish females who were heterozygous for H
and L alleles from females who were homozygous for the H
allele. We can only deduce from this analysis that nonanemic individuals with greater than 3.3% F cells have at least
one H allele.
Phenotypes and genotypes of SS subjects. Due to preferential survival of F cells in the circulation of SS subjects:
percent F reticulocytes must be used as the index of F-cell
production. Because the mean and range of percent F
reticulocytes in anemic SS individuals9J7is greater than the
mean and range of percent F cells seen in nonanemic AS
and AA adult^,^,^,^^ we can potentially discriminate more
precise FCP phenotypes in SS subjects. We used same sex
SS sib pairs to establish phenotypic demarcations. We
reasoned that if differences between discordant SS males is
attributable to H and L alleles at a locus on the X
chromosome a single boundary should separate the percent
F reticulocytes among discordant male sibs. Similarly, one
or two demarcation boundaries should exist for females
(Bimodal LL v HL or H H Trimodal LL v HL v HH). Figure
3 illustrates that among the 15 same sex sib pairs discordant
for F-reticulocyte levels, discordant SS male sibs are separated by a single boundary between 10% and 12% F
reticulocytes, and females are separated by two demarcation boundaries at 10% to 12% and 24% to 26% F
reticulocytes. These observations are consistent with two
genotypes among SS males (L < 12% and H > 12% F
reticulocytes) and three genotypes in females (LL < 12%,
HL 12% to 24%, and HH > 24% F reticulocytes). The
apparent three genotypes in females suggests that the
alleles are codominant and that the putative X-linked FCP
gene may not be subject to X-inactivation. This is in accord
with the elevated level of F reticulocytes in females compared with males.
Hardy-Weinberg analysis. Comparison of F-reticulocyte
w 32E
28
H
MALE SIBS
w
FEMALE SIBS
Fig 3. Comparison of F-reticulocyte levels among SS sib pairs of
the same sex who have discordant F-reticulocyte levels. Discordant
F-reticulocyte levels were defined in a previous study,17 and are
detailed in the Results. Note that the difference between each male
SS sib pair is separated by a threshold (lower hatched area) at 10% to
12% F reticulocytes. while female SS sib pairs are separated by two
thresholds (upper and lower hatched areas) at 10% to 12% and 24%
to 26% F reticulocytes.
819
levels in SS individuals shows that a smaller proportion of
females have a low percentage of F reticulocytes than do
males (see Fig 2). We calculated the fraction of individuals
in both groups with the low phenotype ( < 12% F reticulocytes). Among SS females, 49% (22/45) had less than 12%
F reticulocytes as compared with 76% (44/58) among
males. A significant difference between males and females
in the distribution of F reticulocytes is not compatible with
classic autosomal inheritance. However, if a major determinant of F-cell production is X-linked, the proportion of
males with an L phenotype (9) should be greater than the
proportion of females with the LL phenotype (q2). Similar
differences in the distribution of G6PD activity among
males and females was one of the initial indications that it
was X-linked.3O
Using the phenotype designations described above and
assuming X-linkage, the frequency of the L allele in our
male SS population is 0.76. Using this value for the L allele,
we calculated the expected frequencies of the three female
phenotypes using the Hardy-Weinberg equation. Observed
and expected numbers were not significantly different
(Observed: LL = 22, HL = 19, H H = 4; Expected: LL = 26,
HL = 16, H H = 3), x2 = 1.51, P = .47. Given our data set
(N = 103), the 95% confidence limits of this chi-square
analysis allow the frequency of the L allele in males to be
between 0.57 and 0.79, if X-linkage is assumed.
If autosomal inheritance is assumed, the frequency of the
L allele in our SS male population is the square root of 0.76
or 0.87. Using this value to predict the expected frequencies
of phenotypes in SS females gave values that were significantly different from those observed (Observed: LL = 22,
HL = 19, H H = 4; Expected: LL = 34, HL = 10, H H = 1).
Chi-square analysis shows that our data are incompatible
with autosomal inheritance: xzsquare = 21.34,P = .OOO1.
Pedigree analysis. If F-cell production in SS disease is
regulated by an X-linked locus, the FCP phenotype in SS or
AS males should not be inherited from their fathers.
Alternatively, if F-cell production is controlled by an
autosomal locus, one would expect that one half of the
males should inherit the FCP phenotype from their fathers.
We examined the parents of 20 SS males with F-reticulocye
levels greater than 12% from 18 pedigrees, and one AS
male with greater than 3.3% F cells (Fig 4). In nine
pedigrees where one parent was L or LL ( < 3.3% F cells)
and the other parent had at least one H allele (>3.3% F
cells), only one of 12 possible father-to-son transmissions of
the H phenotype was observed (top row, Fig 4). Two
pedigrees (bottom row, left, Fig 4) were not informative,
since both parents had at least one H allele. Where only the
mother was available for analysis, five of six pedigrees
indicated that the mother had at least one H allele (bottom
row, right, Fig 4). Although not unequivocal, these data
strongly support X-linked inheritance of F-cell production.
We have no explanation for the apparent father-son transmissions (see arrow), but they may indicate that F-cell
production is also influenced by non-X-linked factors.
Linkage analysis. Formal linkage analysis using RFLPs
associated with known positions on the X-chromosome was
performed because the data described above were most
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
DOVER ET AL
820
1.7 13.7 2 5 7.5
9.2 2.0 1.0 4.5 0.6 3.5 0.6 3.9 0.5 13.9 2 7 4.6 0.8 3.8
YE
YEYYYE?
p 9 ~
25
'
26 27
26
A
14 13 5.0 4.0
15
20 26
15
26
23
24 23
18
$ q v l g v
4
16
17
17
41
15 16
25
A
Fig 4. Pedigrees of males with the H-FCP phenotype (>12%F reticulocytes for SS and >3.3% F cells for AS). The numbers associated with AS
individuals are percent F cells, while the numbers associated with SS individualsare percent F reticulocytes. Note on the top row, where both
mother and father were assayed, 11 out of 12 times the mother was either HL or HH (>3.3% F cells). On the second row, in two pedigrees, both
parents had an H phenotype, or, in six pedigrees where only the mother was available for analysis, the mother was either HL or HH five out of six
times. Transmission of the H allele from mother to son was observed 11out of 12 times (top row) and possibly 19 out of 21 times (total pedigrees).
The arrows indicate two pedigrees where mother-toaon transmission of the H allele is tentatively excluded. Half-filled symbols, AS; solid
symbols, SS.
consistent with an X-linked gene having a significant effect
on F-cell production. We analyzed our data first as discreet
phenotypes, and then as a continuous trait. Note that a lod
score of greater than 2 is considered sufficient for establishing linkage to the X-chrom~some.~~
Linkage analysis treating FCP as discreetphenotypes. Using the discreet phenotypes discussed above (see phenotypes of nonanemic adult and phenotypes of SS subjects),
recombinants can be easily identified. This treatment eliminates minor variations within each phenotype that are
attributable to sampling error or other factors unlinked to
the X-chromosome. Erroneous assignment of phenotype
can easily lead to false recombinants and thus nonlinkage
(low lod score), even when two loci are linked. Because the
probability of obtaining false evidence for linkage is small,%
a significant lod score would provide strong evidence for an
PGK (BGL I)
PGKI.9 (PST I)
DXS-17 (TAQI)
DXS-11 VAQ I)
HPRT(BAM1)
ST-14 (TAQ I)
+ - +
5 + 2 - +
5 + 5 - 5 - -
X-linked gene. Our strategy was to use a family with nine
sibs (six SS females, two AS females, and one AS male) to
screen for a candidate linkage site. This family has been
studied repeatedly over several years and full paternity has
been estab1i~hed.l~
As seen in Fig 5, at least two recombinants were detected between the FCP phenotype and each
of the six markers on the long arm of the Xp-chromosome,
leading to lod scores of less than -2 (data not shown). This
excludes Xq as being a candidate region for the FCP locus.
There was no recombination with the Xp loci DXS 43 and
DXS 41, suggesting that the putative FCP locus is located
on the short arm of the X-chromosome.
We then applied linkage analysis of discreet phenotypes
to the rest of our collection of pedigrees, using the linkage
program LIPED. In six additional SS pedigrees and one AA
pedigree, the mothers were heterozygous for one or more
- - -
+ +
+ - +
- +
+ +
()M
+ + +
+ - -
- - +
- +
- -
=AS
?
+ - +
- +
+ -
?
?
?
?
?
?
?
?
?
?
?
?
q12.2-13.1
q12.2-13.1
q22
q24-26
q26
q28
=SS
Fig 5. Summary of linkage analysis in one SS pedigree (Ba) for six RFLP probes on the long arm of the X-chromosome and two anonymous
RFLP probes (DXS 43 and DXS 41) on the shorc arm of the X-chromosome. The numbers below AS individuals refer to percent F cells and the
numbers below SS individuals refer to percent F reticulocytes. The assignments listed on the top row (H, HH, HL) refer to the FCP phenotypes
given to male and female AS and SS subjects as illustrated in Figs 1to 3 and discussed in the text. The designation for the X-linked probes (I&
column) includes the abbreviation for the restriction enzyme usedto test for linkage. The locationfor each probe on the X chromosome is listed in
the right column. With each probe, the mother's two X-chromosomescan be distinguished by the presence (+) or absence (-1 of the polymorphic
restriction enzyme site. For simplicity, only the segregating maternal chromosome is shown for each offspring. Note that no recombinants are
seen with either Xp probe. None of the long arm sites appear to be linked. Half-filledsymbols, AS; solid symbols, SS.
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
821
ADULT Hb F LEVELS ARE INFLUENCED BY AN Xp GENE
RFLP on the short arm of the X-chromosome and had at
least one H allele (Fig 6). Of 27 informative meiosis
(families BA, M, P, SW, H), only one recombinant (M,
11-3) was identified between FCP phenotypes and markers
previously mapped to Xp22.2. Using allele frequencies
derived from analysis of the male cohort data (see HardyWeinberg analysis, L = 0.76; H = 0.24), we observed a
maximum lod score of 4.6 at a recombination fraction of
0.036 (Table 1).Using the allele frequencies determined to
be acceptable by the Hardy-Weinberganalysis (L = 0.57 to
0.79, the 95% confidence range), we obtained maximum lod
scores between 4.6 and 4.5, at a recombination fraction of
0.036. The regions proximal and distal to Xp22.2 were
either unlinked or weakly linked to the FCP locus (Table 1).
This analysis provides evidence for close linkage between
PEDIGREE A
the FCP locus and Xp22.2. Note the maximum lod score for
SS families (BA, M, P, SW) alone is 3.4 at recombination
fraction of 0.04, and the lod score for a single informative
family (BA) was 2.408 at a recombination fraction of 0.
The number of available pedigrees was insufficient to
allow formal segregation analysis to statistically test our
pedigrees for modes of inheritance. Without independent
tests of modes of inheritance, linkage analysis alone can
lead to false assignment of an otherwise unconfirmed
putative ge11e.3~9~~
Methods for correcting lod scores in
analyses where the mode of inheritance is unknown have
been suggested.”~~~
Essentially, the maximum lod score is
not significant unless it is greater than the accepted lod
score plus the log of the number of polymorphic sites tested
while searching for linkage. Ott” also points out that such
PED IGREE P
PEDIGREE BR
PEDIGREE CR
I
I
a FcJl
I
-
+ I - DXS143
+I
DXS
- 84
I :::
DXS 452
DXS 85
+-;
+/-
n
11
- 1
DXS 143
DXS 84
1%
8%
+/-
+/-
+/-
+/+
PEDIGREE M
F relic
DXS143
DXS 84
F cdl
Fd.2
-
1%
DXS 452
DXS 85
-
+
+
FQll
Frotic.
DXS143
DXS 84
PEDIGREE SW
I
T8F&
Fd l
TAU
DXS 85
+/-
-I+
Fall
F rdc.
DX8 85
-/+
F cdl
F recia
DXS 462
DXS 85
DXS 84
m
LL
in tu
F cell
F*
;
+-;
13Fl-?%
n
H
PY
I
F*
+/-
DXS 85
1
PEDIGREE H
Fig 6. Pedigreesof six SS families and one AA family.= The
segregation of Xp probes (DXS
84,41,43,16,85,143, and 452) in
familieowheretheseprobeswere
informative are shown. Note that
the probesat Xp22.2 (DXS 43,16,
and 85) show only one recombination (pedigree M; 11-3) out of
27 informative meioses (pedigree P 11-12: pedigreeM 11-2,3,4,5,
111-1,2,3; pedigree SW 11-1,2,3,4;
pedigree H 11-2,3, 111-2.3.5. and
pedigree BA [Fig 51 11-1,2,3,4,
5,6.7,8,9). Empty symbols, AA;
half-filled symbols, AS; solid
symbols, SS.
rl
I
F call
-I+
LOCATION
xp22.3
Sl%
-I+
I
PROBES
DXS 452
F cdl
+I+
+I+I-
4-I-
DXS 41
xp21.1
4-1.
4-
--
FWH
DXS 85
DXS 16
DXS 43
I
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
DOVER ET AL
a22
Table 1. Two-Point Linkage Analysis Between Variation of F I d l Production and Marker Loci From Four Regions on XP
Lod Score (2)at RecombinationFraction (0)
Location
Discrete phenotypes
Xp22.3
xp22.2
xp22.1
xp21.1
Continuousdistribution
Xp22.3
xp22.2
xp22.1
xD21.1
Locus
.005
DXS 143
DXS 452
DXS a5
DXS 43
DXS 41
DXS a4
-4.12
.43
.a7
.72
.30
4.07
4.22
3.22
2.03
.a2
-3.76
-10.49
.74
-2.02
1.14
-.66
.95
-.19
- .03
-2.23
.44
.72
.56
.22
2.54
3.02
2.36
1.51
.60
-3.21
-11.35
.29
-3.04
.79
-1.24
.73
- .4a
.37
-.13
DXS 143
DXS 452
DXS 85
DXS 43
DXS 41
DXS a4
.10
correction is only necessary if the number of markers tested
is greater than 100. Because we only used 16 polymorphic
markers, Ott’s correction is not required to ensure the
validity of the test. Applying this conservative standard, our
analysis requires a maximum lod score of 3.2 (2 + log 16) to
provide evidence for X-linkage, and thus the lod of 4.6
remains significant.
Linkage analysis of FCP as a continuous trait. Linkage
analysis was also performed after entering the actual value
of F-cell or F-reticulocyte percentage into LIPED, rather
than using discrete phenotypic categories. This treatment
avoids arbitrary assignment of each individual to a particular phenotype and requires only the means -C SD expected
from each phenotype. Using the means f SD estimated
from the female population of our sib pair (see Materials
and Methods) study, the maximum lod score was 3.1 at a
recombination fraction of 0.05 for linkage between the
putative FCP locus and Xp22.2 DNA markers. Since this
lod score is greater than the usually accepted value of 2, it
also provides significant support for X-linkage. The lod
score from regions other than Xp22.2 were again negative
or insignificant. When dealing with continuous traits, the
maximum lod score is expected to be lower than the one
calculated from discreet phenotype, since variations within
each peak of the distribution lower the significance of each
nonrecombinant identified. Our analysis has the added
disadvantage that the means 2 SD used were calculated
from relatively small sample sizes within the total test
population (N = 98). Nevertheless, the result of this analysis is compatible with our previous assumptions and provides evidence for linkage of the FCP locus to Xp22.2.
DISCUSSION
Because Hb F interferes with the polymerization of Hb S ,
S S individuals with elevated levels of Hb F appear to have
less-severe d i ~ e a s e .Indeed,
~ ~ , ~ ~a recently completed natural history study of S S subjects in the United States shows
an inverse correlation between percent Hb F and vasoocclusive crises, which is continuous down to levels of
approximately 4% Hb F?7 Therefore, it has been of
considerable interest to determine what genetic mecha-
.20
.30
.40
,2
(e)
.46
nisms control the diverse levels of Hb F seen in S S disease.
Deletions of large portions of the P-globin gene complex
(deletion HPFH) and point mutations in the promoter
region of the y-globin genes (nondeletion HPFH) have
been shown to increase Hb F levels. However, these
mutations are rare and cannot account for the wide
variation seen in the general SS p o p ~ l a t i o n .Several
~~
investigators have suggested that Hb F levels may be linked
to particular @-globingene haplotypes as determined by
RFLP analysis.lOJ1The Senegal haplotype has been associated with elevated Hb F levels in African, Saudi, and Indian
populations of S S patients. However, when Miller et a1
analyzed Hb F production in individual pedigrees of SS
subjects from the Eastern Saudi Arabia, they noted that
elevated Hb F production could not be explained solely by
association with the Senegal-like haplotype.39In our analysis of haplotypes in the Jamaican S S sib and cohort groups,
we find that less than 10% of the variation in Hb F levels or
percent F reticulocytes is associated with a particular
@-globinhaplotype. This undoubtedly is due to the relative
paucity of homozygotes for the less frequent haplotypes
(Central African Republic and Senegal) in Jamaican and
American S S patients. Our earlier data suggest that 70% of
the variation in Hb F levels in SS disease is attributable to
the variation in percent F reticulocytes, and that, as
described above, the percent F reticulocytes is at least
partially controlled by the FCP locus on the X-chromosome. Further studies comparing the influence of H and L
alleles of the FCP locus in patients with different haplotypes will have to be done to see whether factors associated
with different chromosome 11haplotypes affect the expression of the FCP locus.
The mechanism by which the FCP locus modulates F-cell
production is unknown. Trans-acting proteins bind to enhancer and promoter regions in the @-globincomplex and
may modulate differential globin gene expression. Indeed,
Zon et a P have isolated a DNA-binding protein, GATA-1,
which may serve as one of the major regulators of erythroid
gene expression. The gene for this protein has been
localized to the X-chromosome, at Xp21.1. Another factor
that influences erythroid differentiation in vitro, erythroid
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
ADULT Hb F LEVELS ARE INFLUENCED BY AN Xp GENE
potentiating activity, has been localized to Xp11.4-11.1.41
Since we have clearly defined families that show nonlinkage
(negative lod scores) of the FCP locus to the Xp21.1-cen
region, our data suggest that the FCP locus represents a
gene on the X chromosome which is separate from the two
previously discussed genes. Furthermore, recent experiments by Zitnik et a1,42using hetero-specific hybrids made
from human lymphoid cells fused with MEL cells, are
compatible with the possibility that the gamma gene on
human chromosome 11 may be reactivated by genetic
material on chromosomes other than chromosome 11.
Characterizationof the FCP locus itself will be necessary to
determine whether it encodes for a protein that also binds
to the globin region or, alternatively, affects F-cell production without directly interacting with the P-globin gene
complex on chromosome 11.
Our data suggest that the X-linked FCP alleles are
codominant and not subject to X-inactivation. The mean
F-cell and F-reticulocyte levels are higher in females than in
males (Figs 1 and 2), while the F-reticulocyte level in HL
females is approximately half of that seen in HH females
(Fig 2). In addition, sib pair analysis (Fig 3) indicates the
presence of three phenotypes in females and just two in
males. Although most X-linked genes are subject to dosage
compensation in females, an increasing number of X-linked
genes are being described that escape X-inactivati~n.“~~
These genes exist on both the long and short arm of the
X-chromosome, indicating that escape from X-inactivation
is not a local phenomenon. Currently, four areas of Xp are
known to escape X-inactivation. Two of these, the distal
pseudoautosomal region and ZFX, bracket Xp22.2, the
presumed site of the FCP locus. Brown and WillardMhave
speculated that much of the short arm of the X-chromosome was a recent addition to the eutherian X-chromosome, and suggested that genes throughout its length might
escape X-inactivation.
We presume that the X-linked factor controlling F-cell
production in nonanemic Japanese adults described by
Miyoshi et all8 is the same as the FCP factor described in
this report. Although Miyoshi et a1 suggested that their
factor was X-linked, their pedigrees could not establish
formal linkage to any region on the X-chromosome. In
discussing their observations, Miyoshi et a1 pointed out that
most of the pedigrees describing heterocellular HPFH in
the literature were compatible with X-linked transmission.18The fact that the FCP locus is mapped to the same
region on the short arm of the X-chromosome in both black
SS families and in a white family, suggests that in humans
there is at least one FCP locus on the X-chromosome.
However, it should be emphasized that our analysis does
not exclude other loci either on the X-chromosome, chromosome 11, or other autosomes which may contribute to the
heterocellular HPFH phenotype.
ACKNOWLEDGMENT
We thank R. Schnur, L. Kunkel, P. Pearson, K. Davis, G. Sack,
and B. Vogelstein for DNA probes. We thank Drs H. Franklin
Bunn and Samuel H. Boyer for their thoughtful advice and
encouragement.
REFERENCES
1. Boyer SH, Belding TK, Margolet L, Noyes AN: Fetal hemo11. Labie D, Pagnier J, Lapoumeroulie C, Rouabhi F, Dundaglobin restriction to a few erythrocytes (F cells) in normal human
Belkhodja CP, Beldjord C, Wajcman N, Fabry ME, Nagel R L
adults. Science 188:361,1975
Common haplotype dependency on high G‘-globin gene expression and high Hb F levels in A-thalassemia and sickle cell anemia
2. Chemoff AI: Immunologicstudies to hemoglobins. 11. Quantitative Precipitin test using anti fetal hemoglobin sera. Blood
patients. Proc Natl Acad Sci USA 82:2111-2114,1985
8:413-421,1953
12. Old JM, Ayyub H, Wood WG, Clegg JB, Weatherall D J
3. Marti Von HR, Butler R Haemoglobin F- und Haemoglobin
Linkage analysis of nondeletion hereditary persistence of fetal
hemoglobin. Science 215981,1982
Az-Vermehrung bei der Schweizer Bevolkerung. Acta Haematol
13. Soummer AM, Testa U, Dujardin P, Guerrasio A, Henri A,
26:65-74,1961
Gazak M, Riou J, Rouchant H, Beuzard Y, Rosa J Genetic
4. Serjeant GR: Fetal haemoglobin in homozygous sickle cell
regulation of y gene expression: Study of the interaction p-thalassedisease. Clin Haematol4109-122,1975
mia with heterocellular HPFH. Hum Gene 57:371,1981
5. Zag0 MA, Wood WG, Clegg JB, Weatherall DJ, OSullivan
M, Gunson H: Genetic control of F-cells in human adults. Blood
14. Gianni AM, Bregni M, Cappellini MD, Fiorelli G, Taramelli
53:977,1979
R, Giglioni B, Comi P, Ottolenghi S: A gene controlling fetal
6. Dover GJ, Boyer SH, Pembrey M E F-cell production in
hemoglobin expression in adults is not linked to the non-aglobin
sickle anemia: Regulation by genes linked to p-hemoglobin locus.
cluster. EBMO J 2921-925,1983
Science 211:1441,1981
15. Martinez G, Colombo B: A new type of hereditary persis7. Mason KP, Grandison Y, Hayes N,Serjeant BE, Serjeant
tence of foetal haemoglobin: Is a diffusible factor regulating
GR, Vaidya S, Wood WG: Post-natal decline of fetal haemoglobin
y-chain synthesis? Nature 252:735,1974
in homozygous sickle cell disease: Relationship to parental Hh F
16. Jeffreys AJ, Wilson V, Thein SL, Weatherall DJ, Ponder
levels. Br J Haematol52455,1982
BA: DNA “fingerprints” and segregation analysis of multiple
8. Milner PF, Leibfarth JD, Ford J, Barton BP, Grenett HE,
markers in human pedigrees. Am J Hum Genet 39:11,1986
Garver F A Increased Hb Fin sickle cell anemia is determined by a
17. Boyer SH, Dover GJ, Serjeant GR, Smith KD, Antonarakis
factor linked to the BSgene from one parent. Blood 63:64,1984
SE, Embury SH, Margolet L, Noyes AN, Boyer ML, Bias WB:
9. Dover GJ, Boyer SH, Charache S, Heintzelman K Individual
Production of F cells in sickle cell anemia: Regulation by a genetic
variation in the production and survival of F cells in sickle-cell
locus or loci separate from the p-globin gene cluster. Blood
disease. N Eng J Med 2991428,1978
641053,1984
10. Nagel RL, Fabry ME, Pagnier J, Zohoun I, Wajcman H,
18. Miyoshi K, Kaneto Y, Kawai H, Ohchi H, Niki S, Hasegawa
Baudin V,Labie D: Hematologicallyand genetically distinct forms
K, Shirakami A, Yamano T X-linked dominant control of F cells in
of sickle cell anemia in Africa. N Engl J Med 321:880, 1985
normal adult life: Characterization of the swiss type hereditary
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
824
persistence of fetal hemoglobin regulated dominantly by gene(s)
on X chromosome. Blood 721854,1988
19. Ott J: Invited editorial: Cutting a Gordian knot in the
linkage analysisof complex human traits. Am J Hum Genet 46219,
1990
20. Zielke MR, Meny RG, O’Brien MJ, Smialek JE, Kutlar F,
Huisman THJ, Dover G J Normal fetal hemoglobin levels in the
sudden death syndrome. N Engl J Med 321:1359,1989
21. Serjeant BE, Forbes M, Williams LL, Serjeant GR: Screening cord bloods for detection of sickle cell disease in Jamaica. Clin
Chem 20666,1974
22. Aubourg PR, Sack GH, Moser Hw:Frequent alterations of
visual pigment genes in adrenoleukodystrophy. Am J Hum Genet
42:408,1988
23. Dover GJ, Boyer SH: Fetal hemoglobin-containing cells
have the same mean corpuscular hemoglobin as cells without fetal
hemoglobin: A reciprocal relationship between y- and f3-globin
gene expression in normal subjects and in those with high fetal
hemoglobin production. Blood 691109,1987
24. Dover GJ, Boyer SH, Bell WR: Microscope method for
assaying F cell production: Illustrative changes during infancy and
in aplastic anemia. Blood 52664,1978
25. Antonarakis SE, Boehm CD, Serjeant GR, Theisen CE,
Dover GJ, Kazazian HH Jr: Origin of the p S-globin gene in blacks.
The contribution of recurrent mutation or gene conversion or both.
Proc Natl Acad Sci USA 81:853,1984
26. Weeks DE, Legner T, Wheeler ES, Kaufman C, Ott J:
Measuring the inflation of the LOD score due to its maximization
over model parameter values in human linkage analysis. Genet
Epidemiol7:237,1990
27. Ott J: The likelihood method of linkage analysis, in Analysis
of Human Linkage. Baltimore, MD, Johns Hopkins University
Press, 1988,pp 62-64,78-80
28. Rutland PC, Pembrey ME, Davies T The estimation of fetal
haemoglobin in healthy adults by radioimmunoassay.Br J Haemato1 53:673,1983
29. Wood WG, Stamatoyannopoulos G, Lim G, Nute PE:
F-cells in the adult: Normal values and levels in individuals with
hereditary and acquired elevation of Hb F. Blood 46671,1975
30. Motulskey V, Motulskey A G The Human Genetics: Problems and Approaches. New York, NY,Springer-Verlag,1979, pp
202
31. Vilkki J, Ott J, Savontaus ML, Aula P, Nikoskelainen E K
Optic atrophy in Leber hereditary optic neuroretinopathy is
probably determined by an X-chromosome gene closely linked to
DXS7. Am J Hum Genet 48:486,1991
32. Sherrington R, Brynjolfsson J, Petursson H, Potter M,
Duddleston K, Barraclough B, Wasmuth J, Dobbs M, Gurling H:
Localization of susceptibility locus for schizophrenia on chromosome 5. Nature 336:164,1988
DOVER ET AL
33. Risch N Genetic linkage and complex diseases, with special
reference to psychiatricdisorders. Genet Epidemiol73,1990
34. Thompson E A Interpretation of LOD scores with a set of
marker loci. Genet Epidemiol1:357,1984
35. Perrine RP, Pembrey ME, John P, Perrine S, Shoup F
Natural history of sickle cell anemia in Saudi Arabs: A study of 270
subjects. Ann Intern Med 881,1978
36. Wood WG, Pembrey ME, Serjeant GR, Perrine RP, Weathera11 DJ: Hb F synthesis in sickle cell anaemia: A comparison of
Saudi Arab cases with those of African origin. Br J Haematol
45:431,1980
37. Platt OS, Thorington BD, Brambilla DJ, Milner PF, Rosse
WF, Vinchinsky E, Kinney TR: Pain in sickle cell disease. Rates
and risk factors. N Engl J Med 325:11,1991
38. Economou EP, Antonarakis SE, Kazazian HH Jr, Serjeant
GR, Dover GJ: The variation in Hb F production among normal
and sickle cell adults is not related to nucleotide substitutions in
the y promoter regions. Blood 77:174,1991
39. Miller BA, Salameh M, Ahmed M, Olivieri N, Antognetti G,
Orkin SH, Huisman THJ, Nathan D G Analysis of hemoglobin F
production in Saudi Arabian familieswith sickle cell anemia. Blood
70716,1987
40. Zon LI, Tasi S-F, Burges S, Matsudaira P, Bruns GAP,
Orkin SH: The major human erythroid DNA-binding protein
(GF-1): Primary sequence and localization of the gene to the X
chromosome. Proc Natl Acad Sci USA 87668,1990
41. Huebner K, Isobe M, Gasson JG, Golde DW, Croce CM:
Location of the gene encoding human erythroid potentiating
activity to chromosome region XPll.1-pll.4. Am J Hum Genet
38:819,1986
42. Zitnik G, Hines P, Stamatoyannopoulos G, Papayannopoulou T Murine erythroleukemia cell line GM 979 contains factors
that can activate silent chromosomal human y-globin genes. Proc
Natl Acad Sci USA 88:2530,1991
43. Schneider-Gadicke A, Beer-Romero P, Brown LG, Nussbaum R, Page D C ZFX has a gene structure similar to ZFY, the
putative human sex determinant, and escapes X-inactivation. Cell
57:1247,1989
44. Brown CJ,Willard H F Localization of a gene that escapes
inactivation to the X chromosome proximal short arm: Implications for X inactivation. Am J Hum Genet 46273,1990
45. Fisher EMC, Beer-Romero P, Brown LG, Ridley A, McNeil
JA, Lawrence JB, Willard HF, Bieber FR, Page DC: Homologous
ribosomal protein genes on the human X and Y chromosomes:
Escapes from X inactivation and possible implications for Turner
syndrome. Cell 63:1205,1990
46. Giacomelli F, Wiener J, Kruskal JB, Pomeranz JV, Loud
A V Subpopulationsof blood lymphocytes demonstrated by quantitative cytochemistry. J Histochem Ctyochem 7128,1971
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
1992 80: 816-824
Fetal hemoglobin levels in sickle cell disease and normal individuals
are partially controlled by an X-linked gene located at Xp22.2
GJ Dover, KD Smith, YC Chang, S Purvis, A Mays, DA Meyers, C Sheils and G Serjeant
Updated information and services can be found at:
http://www.bloodjournal.org/content/80/3/816.full.html
Articles on similar topics can be found in the following Blood collections
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml
Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American
Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.